Abstract
This article presents an argument for how to generalize from a single case study evaluation of a social programme or policy, focusing on two essential elements of case study – context and particularity. Taking an interpretivist perspective and drawing on artistic and humanistic ways of understanding, it examines both different ways of generalizing from the case that retain a connection with the context in which they first arose and how we generalize by direct encounter with the particular. This argument is not new. It has its origins in earlier centuries and, in contemporary evaluation, dating back to the 1970s. However, in a political climate that privileges evaluation approaches stemming from large sample studies and experimental designs, it seems timely to restate the value of generalizing from the single case. Grounded in the reality of programme experience and retaining that connection with context and particulars in the case facilitates the use of evaluation knowledge.
Keywords
Introduction
This article has a single theme and purpose – to demonstrate how one can generalize from the single case in case study evaluation by focusing, in particular, on two essential elements of case study – context and particularity. My theoretical approach, outlined further in the next section, is interpretivist and constructivist. Generalizations from case study evaluations conducted within this framework not only retain a connection with the context in which they first arose, they also invoke the agency of the person generalizing from them.
Case study evaluation can be conducted from other theoretical perspectives, of course, such as post-positivist or theory-led, incorporated in different evaluation designs – experimental, impact and mixed method, for example, and reach generalizations in different ways, such as with QCA techniques or other analytic processes (Gomm et al., 2000; Small, 2009; Yin, 2013). These forms of generalization are all legitimate for the purposes these different approaches serve. However, they are not my emphasis here. My focus is on how we learn from the singular case and the generalized understanding it promotes, taking an interpretive and constructivist perspective of reality.
It seems timely to review the usefulness of case study evaluation from this perspective for its intrinsic worth has been overshadowed in current times by the dominance of methodologies, such as experimental, quasi-experimental and survey, utilizing quantitative outcome measures of programme effectiveness that promise greater ‘certainty’ and demonstration of impact on limited measures of worth. The inadequacy of such methodologies for informing future development and policy of social and educational programmes has been known for decades. It is one of the reasons why case study evaluation evolved in the seventies to generate more in-depth understanding of complex programmes and the people who enact them in particular socio-political contexts (Simons, 1980, 2009). However, this epistemological shift in how we come to know and understand such programmes seems to be forgotten, ignored or side-lined in current methodological preferences for evaluation of social programmes. This point was reiterated by Lincoln and Canella (2004) and Lather (2004), some 10 years ago and has been emphasized, most recently by Patton (2014) in challenging the ‘so called’ gold standard of randomised controlled trials in evaluation studies. In recent years, of course, case studies have been utilized in mixed method evaluation studies (e.g. Greene, 2007; Stern et al., 2012), but here I am only discussing single case study evaluation.
There is a further reason for rekindling an interest in single case study evaluation and how we generalize from it at this point in time. In the unstable social and political world of today where conditions are changing so rapidly, we may be missing an opportunity to understand in-depth how programmes and policies are being interpreted and the impact these have in different contexts and circumstances. And this may have ethical and political consequences for those traditionally not seen, heard, or well served in many societies and cultures. We need their stories in order to identify the key issues to evaluate and reveal the conditions in which policies are enacted to argue more strongly for policy development that would make a difference to their lives. Dominant models in currency, such as those indicated above, have advantages of scope and scale, but they cannot capture the experience and insight of programmes in action and apprehend the consequences for people that case study evaluation is designed to do.
Narrative background and rationale
My argument stems from my history of conducting case study evaluation over 35 years and what I have learned. This has meant leaving behind my roots and training in experimental psychology. Making sense of cases in programme and policy evaluation for me draws more on an integration of concepts from different disciplines, such as social and humanistic psychology, politics, history, anthropology, biography and the sociology of groups than concepts drawn from experimental psychology. It also relies on intuitive modes of interpretation and understanding as much, if not more, than explanation. These are not discrete processes. Both may be important in coming to an understanding of the case. But my leaning is towards ‘the art of interpretation’, the key process in the social sciences for making sense (Denzin, 1994: 500).
My understanding of how to generalize from single case study evaluation parallels this route and connects with my English literature background and training as a teacher and psychologist. Over the years I have found I have learned less from propositional forms of knowledge than I have from narratives, critical incidents and stories of people in context and circumstance. In aspiring to reach audiences to act upon evaluations, I have also discovered such forms of understanding facilitate action as they communicate in ways others can readily comprehend.
So now, my understanding of how to generalize from the single case is closely related to the humanities and the kind of understanding we gain from reading poems, novels and short stories and from the meaning we make of the art we behold. I will elaborate this perspective later and relate this to how others can learn from case study evaluations. First, however, I briefly outline what values and principles underpin my practice.
The theoretical lens through which I view the world and the world of case study evaluation is interpretivist and constructivist, both in a cognitive and social sense (Schwandt, 2001: 31–3). 1 It values how people construct and understand their experience, how their intentions and perceptions affect how they act in specific situations and how they make sense of what they do. In evaluation we cannot rely on ‘subjective meanings’ alone, of course, in making sense of a case. There is a social reality of which individuals are part where other people may have different subjective meanings of the same event or experience, and a political reality that may affect how they interpret the situations they face in any particular case. Single case study evaluations within this perspective are often co-constructions of the reality encountered in the particular programme and the specific socio-political context.
Methodologically, accessing subjective meanings often means taking a phenomenological approach to interviewing but this is balanced and integrated in the cases I evaluate with other methods. These are primarily qualitative, though there are occasions when in-depth understanding of a particular case is enhanced through the inclusion of quantitative methods. Politically and ethically I adopt a democratic perspective to the conduct of the study in which participants, whatever position or power they hold within or beyond the case, are treated equitably in the process of determining the theory and validity of the case. And practically, I am committed to rich portrayals of instance and circumstance and narrative and artistic ways of reporting to provide our audiences with familiar modes of understanding what was learned in the case.
Case study evaluation: Different perspectives
Case study can be perceived in so many different ways that it is important to indicate what concept of case study evaluation I advocate in this article, as it relates to the ways I suggest we can generalize from the single case. We might all agree on a basic definition. Case study is the study of the singular, the particular, the unique, whether that single case is a person, a project, an institution, a programme or a policy. At this point differences emerge. The case may be selected for its intrinsic, instrumental or collective use (Stake, 1995: 3–4). It may be studied methodologically in different ways, as in pre-experimental exploratory studies, as an exploration of a particular theory in different settings, or as one of a number of specific cases to represent regional differences of an innovative programme. It may utilize different methods, qualitative and quantitative, according to which offer the most appropriate evidence for understanding the case. It may be analysed inductively or deductively. These differences in conception and use notwithstanding, what several authors do agree on (e.g. Merriam, 1988; Simons, 2009; Stake, 1995; Thomas, 2011; Yin, 1994) is the importance of understanding in a ‘real life’ context.
Case study evaluation may also be presented in different ways, as a reported story of a single case, as a summary or conclusion from a number of cases, as an interpretative account. Other cases may be written to include detailed evidence from which it is possible to see how understandings or conclusions were reached. This is most evident in historical and documentary case study evaluation. With such a range of possibilities of how to construct and conduct case study evaluation, there are clearly choices to be made, theoretically, epistemologically and methodologically.
The kind of in-depth case study evaluation that forms the basis for discussion of how to generalize from the single case in this article is one that, whether chosen for its intrinsic interest or instrumental use, is theory-generated rather than theory-led, is analysed and interpreted inductively, employing in the main, qualitative methods to understand the actions and circumstances of the case in the contexts in which they occurred.
The case will be richly described and evidence-based, in the form of observations and perspectives of stakeholders and participants, significant incidents, narratives and critical analysis of any relevant documents. The primary purpose is to generate in-depth understanding of the value of programmes and policies to inform policy development, professional practice, and civil and community action.
The centrality of context
To say ‘context matters’ is often taken for granted in case study evaluation. But as it is central to my argument for how to generalize, let me elaborate a little. Context is not simply a location or static description of time, place and circumstance. In the overall ‘programme case study context’ there will be several conceptions and levels of context: a cultural context infused with different norms and assumptions (Hyatt and Simons, 1999); a political context inclusive of different interests and values; a person context of key people and roles; a subject context, its history and focus; and a policy context to explore how and why this particular programme or policy emerged at this time. Beyond these specific contexts, there will be a broader societal context which may constitute an external influence or pressure on the case. For further exploration of the different conceptions of context and how they may be used see Rog et al. (2012).
Not all these contexts maybe relevant in any particular case. The point is that in making sense of the case and reporting findings, it is important to recognize how these different levels of context interact to produce the meaning or theory of the case. Through the interweaving of these contexts, as well the particulars in the case, the case study evaluator aspires to represent the dynamics within and between these levels of contexts to arrive at a holistic understanding. It is this, if well portrayed, preferably visually to get beyond linear description, which enables users of case study evaluations to generalize to their own contexts or apprehend the generality of the findings.
Interpret in context: Generalizing from the single case in evaluation
If we take a leaf from the OED, it is clear that generalization can be conceived differently. But essentially, to generalize is how to make inferences or make a general statement about the likelihood that a finding established in one case or setting has significance in another or to another person. In case study evaluation, generalization can take many forms: formal or informal, analytic or interpretative, from a sample to a population, from within the case itself. So there are choices to make about how best to generalize for the specific purpose the case study evaluation hopes to serve. Policy makers, for example, may prefer formal generalizations that aspire to prediction across a number of cases, whereas practitioners may prefer to engage more informally with how they generalize from the case itself to their own setting.
My emphasis is how to generalize from the single case taking an interpretive perspective, portraying the case in its particularity and with sufficient context to enable others, whether policy makers or practitioners, to engage with the issues in the process of generalizing to their own or similar contexts or for their own purpose. Generalizing from multiple cases is not my prime focus nor is propositional generalization. Analytic generalization (Gomm et al., 2000; Small, 2009; Yin, 2013) I also leave to others, as it stems from a different epistemological perspective and ways to generalize.
In naturalistic inquiry, which is closer to my perspective, ‘transferability’ is often suggested as an alternative to generalization. This is a term introduced by Guba and Lincoln (1985) in articulating parallel criteria for qualitative inquiry to traditional criteria of internal and external validity. I prefer not to use this term for it does not connect well with the process of generalizing I am advocating and the agency of the person making the generalization. Guba and Lincoln themselves realized its shortcomings when several years later they indicated that the parallel criteria they had outlined for qualitative inquiry retained a positivist ring (Guba and Lincoln, 1989: 245), and they went on to argue for the concept of ‘authenticity’ as a more appropriate standard for determining the worth of qualitative inquiries (see also Lawrence-Lightfoot and Davis, 1997: 14).
My starting point for thinking about generalization in case study evaluation was triggered by a paper by Robert Stake in which he argued for a form of generalization termed naturalistic, which draws on the tacit knowledge we have that enables us to recognize similarities and differences in the case in order to generate new understanding. This he opposed to ‘formalistic generalization’, which is more concerned with explanation, with formalising propositional statements of law-like relationships which philosophers of the positivist school, he states, claim are the ‘closest approximations of Truth’ (Stake, 1978: 6).
‘Naturalistic generalization’ on the other hand relates to understanding and the claims made by philosophers of anti-positivist persuasion that: Truth in the field of human affairs is better approximated by statements that are rich with the sense of human encounter. To speak not of underlying attributes, objective observables, and universal forces, but of perceptions and understanding that come from immersion in and holistic regard for the phenomena. (Stake, 1978: 6)
While case study may be used by those who search for explanatory laws and building theory, Stake was emphasizing here that its appeal is more to experiential and humanistic understanding and has in fact ‘an epistemological advantage over other inquiry methods as a basis for “naturalistic generalization”’ (Simons, 1980: 62).
My interest in generalizing from the single case in evaluation was further stimulated by Cronbach’s (1975) paper, ‘Beyond the two disciplines of scientific psychology’, in which he argued that the two disciplines of scientific psychology (experimental and correlational research) could not serve contemporary needs for knowledge of social programmes, as they failed to account for contextual factors of time and place and the uncontrolled nature of complex social situations. His proposal, in indicating a need for descriptive studies, was to reverse our priorities. Instead of making generalization the ruling consideration in research, he suggested we should observe and interpret effects in context: As he [the observer] goes from situation to situation, his first task is to describe and interpret the effect anew in each locale, perhaps taking into account factors unique to that locale . . . That is, generalization comes last, and the exception is taken as seriously as the rule. (Cronbach, 1975: 125)
Cronbach’s elaboration of this position Hamilton has observed ‘provides a concise rationale for case study inquiry’ (Hamilton, 1980: 87).
More recently what gave a fillip to my interest in how we generalize from the single case was the discovery of a text by Reynolds (1908), which illustrates that generalization from the particular is not new. The significance of this point is explored in a later section. For the moment let me state the different ways in which we can generalize from single case study evaluation, staying close to the context and authenticity of events, circumstance and action. I first focus on the contextual element in generalizing and, second, address how we generalize from direct encounter with the particulars in the case. In reality, of course, the two are interconnected.
Generalizing in case study evaluation connected to context
The forms of generalizing outlined in this section all differ from the kind of generalizations adopted in research that utilizes large samples and quantitative measures, making inferences from data stated as formal propositions and which aspire to predict from a sample to a population. They also differ from those in smaller samples which seek to determine causality for the effects that are observed
Making inferences from cases with a qualitative data set arise more from a process of interpretation in context, appealing to tacit and situated understanding for acceptance of their validity (see Flyvberg, 2006; Simons et al., 2003; Stake, 1978, 2006). Such inferences are possible where the context and experience of the case is richly described with sufficient detail of time, place and person for the reader to recognize and connect with the events and experiences portrayed. Generalizations then are not abstractions, independent of place and context, but depend for their meaning on maintaining a connectedness with the particulars of the concrete case in context. This resonates with Flyvbjerg’s (2006) observation and argument that, in the study of human affairs, ‘there appears to exist only context-dependent knowledge’ (p. 221).
The most familiar form of generalizing in case study evaluation is cross-case generalization, in collective or multi-site case study. Each case is explored to see if issues that arise in one case also exist in other cases and what interconnecting themes there are between them. This kind of generalization has a degree of abstraction and potential for theorizing and is often welcomed by commissioners of evaluation concerned that findings from the single case do not provide an adequate or ‘safe’ basis for policy determination.
Cross case analysis may be achieved deductively, through examining all the cases to see whether the issues which were consistently explored from the outset, often within a particular theoretical frame, are sustained in all cases and how they interconnect to produce themes or patterns that make sense of the programme as a whole. Or it may be achieved inductively by a reading of all the in-depth cases to determine what issues are significant in each and only then exploring whether there is sufficient overlap between them to provide a basis for generalizing or whether different issues make sense in specific cases. At first sight this way of generalizing may appear to be close to propositional generalization, but it is different in that it does not seek to ‘project to cases that are unexamined’ (Stake, 2006: 88). The search is for common understandings between cases while also recognizing their situational uniqueness.
The forms of generalization which follow – naturalistic, situated, concept and process – rely more on the individual making the inferences, drawing both on tacit knowledge (Polanyi, 1967) and in-depth understanding of context to generalize to other cases and places.
Naturalistic generalization, first proposed by Stake (1978), as noted earlier, is reached on the basis of recognition of similarities and differences to cases with which we are familiar. To enable such recognition, the case needs to feature rich description, people’s voices and enough detail of time, place and context to provide a vicarious experience that will help readers recognize what is similar and what is not to their own context (Lawrence-Lightfoot, 2005; Stake, 1978). It is a process of discerning what might and might not be different, not a direct translation from one context to another. Donmoyer (1990: 185) makes a similar point. This is one of the reasons why I do not use the concept of transferability referred to earlier. The process of generalizing is much more subtle than transferability infers. Transferability does not capture the dynamic, iterative and different ways in which we can generalize in single case study evaluation, which is more an immediate apprehension or recognition of what relates to or resonates with the reader’s situation or experience.
Situated generalization (Simons et al., 2003) is close to the concept of naturalistic generalization in relying for its generality on retaining a connectedness with the context in which it first evolved. However, it has an extra dimension in a practice context. This notion of generalization was identified in an evaluation of a research project that engaged teachers in and with research. Teachers engaged in research at several levels: by themselves, in partnership with colleagues in the one school, and in partnership with colleagues in other schools. Their investigations ranged from a case study of a single issue in one classroom to case studies of issues conducted in several classrooms. Both qualitative and quantitative methods were employed. Teachers were also encouraged to engage with research findings from external research. However, the evidence from the evaluation indicated that they were more likely to recognize the significance of generalizations for their own practice stemming from their colleagues than from external researchers. In other words, in addition to the usual technical validity criteria to establish significance of the findings, generalizations were seen to be dependable if trust existed between those who conducted the research (teachers in this example) and those thinking about using it (other teachers).
There may also be a broader generalization at stake here. The fact that teachers share common understandings of what it is like to teach and what happens in schools, as well as subject content in some instances means that the generalization they detect will not be theirs alone, even if they determined it independent of other teachers.
This example highlights the importance of agency in enhancing the use of generalizations, in this case to improve professional practice. It resonates with what Flvyberg (2006) has observed in his research on learning, that working from cases and context-dependent knowledge is the way in which novices or rule-base beginners become experts (221–3). This way of generalizing is applicable in other practice contexts and in policy contexts to give those involved in policy making access to the vicarious experience of the case to inform their decision making.
Concept generalization concerns what you discover in making sense of the case itself and how you choose to frame it. In the process of interpreting and analysing, you begin to generate a theory of the case that makes sense of the whole. Concepts may be identified which, though making sense in the one case, have equal consequence for other cases of a similar kind but in different contexts. It is the concept that generalizes not the specific content, issue or effect. The internal validity of the concept/s is located in the logic of the argument that leads to its significance as a means of understanding what transpired in the case and the data that are evident to support it.
One example here stems from a qualitative study of first year undergraduates’ pre- examination preparations where, through analysis of in-depth interviews with a number of undergraduates, it was possible to determine attitudes to their exam preparation, through an analysis of how they responded to cues from the tutor concerning revision for the exam. Student responses fell into three distinct groups: those who listened carefully and explicitly sought cues from the tutor; those who, despite the tutor spelling out precisely what was required for revision and possible inclusion in the exam, ignored all advice; and those who had no particular interest. The first group were designated ‘cue-seekers’, the second group ‘cue deaf’, the third group ‘cue-indifferent’. These concepts were derived from a study of one cadre of students in one subject, but it is easy to see how they could apply to a different year group and different academic subject.
This concept of generalization is similar to the process Donmoyer (2008) identifies of ‘intellectual generalization’ (quoted by Butler-Kisber (2010: 15) to indicate the cognitive understanding one can gain from qualitative accounts or cases even if settings are quite different.
Process generalization is determined in much the same way as concept generalization. In making cognitive sense of the case, it is possible to identify a key process in one case or several that led to successful outcomes, which translate to other contexts, irrespective of the precise content of the case(s). As with the other forms of generalizing in this section, sufficient context needs to be described to enable others to see whether the process is likely to work in another setting. One example here is the collaborative model for sustainable school self-evaluation identified in researching school self-evaluation in a number of schools and countries (Simons, 2002). Schools that successfully sustained a process of school self-evaluation had an infra-structure that was collaborative at all stages of the evaluation; from the choice of issues for evaluation and the setting of aims and key questions, to the collection and analysis of data, to discussion of findings and implications for action. Here, again, it is not difficult to see that the collaborative strategy deemed successful in this example could equally well be adopted in other organizations that wanted to institute a self-evaluative process to facilitate the operations of the institution. As with concept generalization, it is the process that generalizes not the substantive content or findings from the case.
Particularization
The different forms of generalization discussed above are useful in many formal evaluation situations to validate the significance of findings and persuade others to use them. But the overarching justification for how we learn from case study is particularization – a rich portrayal of insights and understandings interpreted in the particular context. Several authors have made this point (Flyvberg, 2006; Simons 2009; Stake, 1995). Stake (1995: 8) states it strongly when he observes that ‘[t]he real business of case study is particularization, not generalization’, referring to the main reason for studying the singular, which is to understand the uniqueness of the case itself.
The universal in the particular: A paradox
My perspective, explored further in Simons (1996, 2009) and Simons and McCormack (2007), is similar in that I believe the ‘real’ strength of case study lies in the insights we gain from in-depth study of the particular. If we study the singular case in sufficient depth, and are able to capture its essence – what makes it unique – in all its particularity, I believe we will also discover something of universal significance.
This is something of a paradox. The more you capture the particulars of one person, context, programme, policy, its context and circumstance, the more likely you are to discover something universal. This may not be a familiar process or way of understanding in much of the evaluation work we do, yet it can provide an insightful basis for action beyond the specific case (see Abma and Stake [2014], for an excellent example of the potential of this way of thinking in the single case of an individual; and Kushner [2000], for his descriptions and episodes of individuals’ experience in programme case study evaluation).
Seeing the universal in the particular is a powerful way of generalizing; it resonates both with the way in which many discoveries are made in science – ‘The Case of the Midwife Toad’ (Koestler, 1971) is often cited as an example – and in how we learn from artists, poets and novelists who reach us by communicating recognizable truths about human relationships or social conditions. As MacDonald and Walker (1975) have noted: Case study is the way of the artist, who achieves greatness when, through the portrayal of a single instance, locked in time and circumstance, he communicates enduring truths about the human condition. (MacDonald and Walker, 1975: 3)
Lawrence-Lightfoot (2005) has made a similar point in her discussion of portraiture, in noting the paradox recognized by novelists, poets and playwrights that ‘as one moves closer to the unique characteristics of a person or a place, one discovers the universal’ (2005: 12). And in commenting on Allen’s (1995) book, ‘Art is a Way of Knowing’, McNiff (1998: 36) notes that, as Allen records her personal experience, she simultaneously identifies universal themes and principles manifested through her experience. We have no difficulty recognizing this paradox in the work of major artists. Cézanne’s trees provide a perfect example, as May (1994: 77) has observed: ‘Cézanne sees a tree. He sees it in a way no one else has ever seen it’. It is Cezanne’s personal experience of the tree and this is literally new, unique and original. But at the same time, says May, ‘I can say without exaggeration that I never really saw a tree until I had seen and absorbed Cézanne’s painting of them’ (May, 1994: 78).
In other words intensive exploration and engagement, cognitively and intuitively, with the particular, is, or becomes ultimately, ‘a unified concept, a universal understanding’ (Simons, 1996: 240). Just as T.S. Eliot noted at the end of the Four Quartets, ‘And the fire and rose are one’.
Taking a cue from artists and poets
So it may be that by taking the role of the artist rather than the scientist in conducting and reporting case study evaluations, or exploring similarities in how discoveries are made by both, we can reveal the ‘truths’ of the specific case in context that others can immediately apprehend. The scientist in us often seeks to codify, categorize, explain, whereas the artist aspires to reveal what is hidden, challenge what is the norm and engage directly with the object s/he is seeking to portray (Stake and Kerr, 1994). The artist can live with ambiguity, uncertainty, or, in the words of John Keats (1818), ‘negativity capability’, that is, when a person ‘is capable of being in uncertainties, Mysteries, doubts, without any irritable reaching after fact and reason’ (quoted by Wilson, 2008: 131–2).
Staying with the inconclusiveness of the paradox, and having the patience not to rush to judgement, is another way of putting this to see what understanding eventually transpires. Seeley and Reason (2008) make this point well in illustrating that we need to take space and time and use all our senses in exploring experiential and presentational knowing, before rushing to propositional knowing. Exploring the paradox, as I have argued elsewhere, also offers us the opportunity to discover something new. ‘To live with ambiguity, to challenge certainty, to creatively encounter, is to arrive, eventually, at “seeing” anew’ (Simons, 1996: 238). It is akin in the world of evaluation case study to getting lost in the data in order to learn something we did not know or something we did know but come to understand differently.
Artists often comment that it is not for us to try to interpret their meaning; we should see what we see. Magritte, for instance, wrote of his paintings, ‘It does not matter what my paintings will be worth in a hundred years’ time or that people will see what I see. What matters is that they see differently’ (Magritte, 1998). Taking a cue from Magritte here, in case study evaluations we should acknowledge the way in which we arrive at our ‘best’ interpretation of the case but leave space for others to ‘see’ it differently.
Poetic form
While artistic forms of all kinds are useful to give us a different lens on how we perceive and report the experience of the case, poetry, I find, is particularly helpful for gaining and sharing insights from the case. For poets are also trying to share insights of particular human actions, circumstances and times. And they can do this frequently more succinctly and powerfully than we can. The analogy with poetry resonates strongly because of its closely observed detail, its capacity to capture the ‘essence’ of the subject of the poem and the intensity and simplicity of its form. Richly observed detail, interpreting in context and portraying the essence of the particular are all characteristics of case study evaluation, though we still have much to learn about how to communicate. The poetic form is powerful for this purpose. Witness Laurel Richardson’s (1997) now well-known example of ‘Louisa May’s Story of Her Life’, where Richardson transposes the experience Louisa May talked about in an in-depth interview into poetic and dramatic form using Louisa May’s exact words, syntax and grammar. The authenticity and agency of Louisa’s experience is retained and gives readers a heartfelt insight into that experience.
When we reflect how often it is that we take a leap of understanding or gain an ‘aha!’ signal from a poem or fragment of a poem that forces us to think again, we might do well to consider using the poetic form more often in our evaluations to communicate the universal in the particular.
From the particular to the general is not a new issue 2
From time immemorial, certainly since ancient Greece (Stake, 2006: 88), there has been a tension between the particular and the general. This is still evident today in the choices different authors make on whether to generate theory from the case(s), formulate general propositions across cases, or stay close to the particulars within the case itself. It is an epistemological question as much as a procedural one, as Stake has pointed out (2006: 7). This poses a dilemma for multi-site case study, he recognizes, as these are more usually studies of particularization than generalization (2006: 8). He addresses this dilemma, where audiences are seeking to generalize from more than one case, by designing a study with several cases, but with the intent to understand the whole through asking some overarching questions. However, both in his 2006 book and an earlier multi-site case study evaluation (Stake and Easley, 1979) he draws our attention again and again back to the point that much of the meaning is in the situationality or particulars of the case: ‘the power of the case study is its attention to the local situation, not how to represent other cases in general’ (Stake, 2006: 8).
That generalizing from the single case itself is far from new, is illustrated in the quotation below from a Preface to a book, A Poor Man’s House, written in 1908.
3
Stephen Reynolds, the author, was studying male poverty. From recorded notes in a journal and in letters to a friend, he was going to write fiction but decided that ‘real life’ and people were far more ‘telling’ than anything he could invent. He aspired to present a typical case of a poor man’s house and life, but noted that the conclusions possibly went beyond this one man’s life. So he offered the following explanation of how he reached his conclusions: Educated people usually deal with the poor man’s life deductively; they reason from the general to the particular; and, starting with a theory, religious, philanthropic, political, or what not, they seek, and too easily find, among the millions of poor, specimens – very frequently abnormal – to illustrate their theories. With anything but human beings, that is an excellent method. Human beings, unfortunately, have individualities. They do what, theoretically, they ought not to do, and leave undone those things they ought to do. They are even said to possess souls – untrustworthy things beyond the reach of sociologists. The inductive method – reasoning from the particular to the general . . . should at least help to counterbalance the psychological superficiality of the deductive method . . . . (Reynolds, 1908: Preface)
This may be overstated but it is a useful reminder that, in our search for formal generalizations abstracted from context, we can lose contact with the concrete situations from which they arose. Not only does this fail to capture the uniqueness and humanity of individuals, it reduces them in the process, failing to present their experience in any ‘real’ sense. Going even further back, William Blake, in Annotations to Sir Joshua Reynolds’s Discourses, decidedly chose to value the particular over the general when he said ‘To Particularize is the Alone Distinction of Merit’ (c. 1798–1809: xvii–xcviii). And in ‘The Marriage of Heaven and Hell’, he reinforces this perception, as Wilson (2008) points out, noting that favouring the abstract over the concrete, one ‘sees all things only “thro” the narrow chinks of his cavern’ (Wilson, 2008: 62).
The danger Wilson is pointing to here is that abstraction relies heavily on what we know from our past understanding of things and this may prevent us experiencing a concrete event directly or ‘apprend[ing]’ a particular moment (Wilson, 2008: 63). Blake fought hard, says Wilson, to break through mental barriers to something unique and living. It is this search for the unique and living we should take from the Blake example to remind ourselves of the possibility of discovering something ‘new’ beyond our current understanding of the way things are.
Even if we did not have these examples, from centuries ago, we can point to the last century and this when, surprisingly perhaps, policy has sometimes been influenced not by the conclusions of randomized controlled trials or other evidence determined by ‘scientific’ means, but by what policy makers know about education or social care, through their personal contacts, for example, their granddaughter’s schooling or their own, or what has happened in social life through the deaths of neglected children or the triumphs of particular people over adversity. Particular stories often hold the key to why policies have not worked well in the past or how they may offer hope for the future.
In other words people have been generalizing from the particular for quite some time. What is significant about this way of generalizing is the immediate connection to the humanity of individuals in the case, the reality of the context, the ‘lived experience’. It provides an authentic context and illustration which people recognize and from which they can learn.
Portraying and learning from single case study evaluation
How one learns from any evaluation to advance policy or practice has much to do with how it is conducted and presented, its validity and credibility, but also, critically, with how people choose to learn or the assumptions under which they think their organizations learn (Rogers and Williams, 2006). This is no less true of case study evaluation. Some may prefer aggregated knowledge displayed in the case in narrative, graphic and summary form, including conclusions and explanations, others ‘thick descriptions’ (Geertz, 1973) of events and circumstances in the complexity of the socio-political context. Yet others, a highly interpretative account. The preference may relate to the purpose for which the evaluation was conducted and the audience one hopes to persuade.
For the kind of case study advocated in this article, case study evaluation has huge potential for those who choose to learn from humanistic ways of understanding as indicated in the section taking a cue from artists and poets. For the case can be communicated in ways that are readily accessible through portrayals of people, experience and context and in language and forms, such as documentary, narrative and story, that are familiar ways of understanding. This increases the likelihood that the findings will resonate and be acted upon.
To enable this to happen, the case, context and significant issues discovered in it, need to be richly described to enable people to generalize in the ways described earlier on the basis of recognition to other contexts and places. It should include closely observed episodes, vignettes, excerpts from interviews, observations, cameos of individuals, critical incidents, short stories. It may include data derived from quantitative measures – graphs of patterns of results, for example, if they add to the understanding of the case, and, depending upon the ethical clearance possible, visual forms such as photographs, video diaries or video clips. These different forms need to be skillfully interwoven grounded in the reality of what transpired to tell the story of the case. Given our digital age, there is further scope for presenting and communicating complex evidence from case study evaluation embedded in context in short episodes, short stories and short films.
In recent years, there has been a growing interest in using artistic forms – poetic, dramatic, collage, drawing – in the gathering and analysis of data and in reporting findings ( Butler-Kisber, 2010; Liamputtong and Rumbold, 2008; Seeley and Reason, 2008; Simons and McCormack, 2007). These have often been adopted in evaluations in professional practice contexts of education and health care. Yet they are potentially relevant in evaluation policy contexts too, for the universality of the arts offer all people the opportunity to see and understand differently. It might be more of a challenge in policy contexts to accept the uncertainty that such forms of understanding promote. However, it does gives agency to those in positions of responsibility to engage with the issues in the case to inform actions and policy deliberations.
Narrative structure and story
Whatever form of presentation is adopted – written, audio or visual, and I appreciate that in some policy contexts we may be obliged to report more conventionally – it is paramount for the case to have a strong narrative structure. As House (1980: 102) has advised: ‘Every evaluation must have a minimum degree of coherence. The minimum coherence is that the evaluation tells a story.’ How one arrives at this story and how one tells it can vary considerably, which again provides scope for how different people learn.
Finding that strong narrative structure in case study evaluation is akin to establishing a theory of the case, that is, the story that demonstrates the value of the programme or policy. 4 This theory is determined at the end of the process of iterative analysis and interpretation, and interweaving of data to reach coherence. It is similar to the way in which the documentary film maker, Frederick Wiseman, reduces many hours of filming to a two or three hour film, ordering and reordering frames and episodes until a story is constructed that makes coherent sense. It is a theory he arrives at, he says, rather than one with which he began (quoted in Graham, 1971: 44). The theory of the case can be told in many ways, and formal narrative or documentary forms are most common.
Telling the theory of the case in story form is quite different. However, providing the underlying narrative structure and warrant for validity is clear and the story is told well, it offers people of many persuasions the opportunity to learn from the single case. Stories, after all, according to Ben Okri (1997) are the natural way in which we learn.
Final words
This article has outlined several ways in which we can generalize from the single case in case study evaluation and has ended with a focus on particularization and how we can learn from it. However, there is still more we can do to strengthen and extend ways to generalize from the single case. More narratives is one way forward, as it is through people that cases are proposed, enacted and evaluated. These should not only be narratives of participants but also those who generate policies and programmes, who commission the evaluation and of evaluators themselves to illustrate the power and influence they each had in establishing the value of programme or policy and how their values impacted on the findings. These would all need to be sensitively portrayed, especially when key protagonists and volatile socio-political contexts might be identified.
A second development of equal importance is to find ways to portray the different levels of contexts in the case and how they interweave to show the interrelatedness of programme and policy implementation. This is crucial to enable people to ‘interpret in context’ without reducing the complexity. Case studies are often presented in a linear format that does not convey how these levels of context intersect and influence each other. The use of digital media could help raise the game here. And, taking a leaf from the poets and artists who also evaluate the social world and conditions, we could further explore artistic processes in the conduct and analysis of cases as well as in their portrayal. All these developments could enhance opportunities for people to generalize from the single case, related to what best serves their evaluation purpose and how they learn.
And there I rest my case.
Footnotes
Acknowledgements
I am grateful to three anonymous reviewers for their generous comments on this article.
Funding
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.